hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mahmood N <nt_mahm...@yahoo.com>
Subject Re: ipc.client RetryUpToMaximumCountWithFixedSleep
Date Tue, 05 May 2015 20:12:58 GMT
Finally I figured out the problem! Things are:
1- What are specified in the command as ADDRESS and PORT
java -jar indexdata.jar result hdfs://ADDRESS:PORT/data-Index
must be the same things as fs.default.name<property>
  <name>fs.default.name</name>
  <value>hdfs://ADDRESS:PORT</value>
</property>


So, having <value>hdfs://hostname:9000</value> and running hdfs://localhost:54310
fails to work (ipc.clinet time out). Even the hadoop report command says namenode is up because
it read the fs.default.name


2- After matching ADDRESS and PORT, I got this errorException in thread "main" java.lang.NullPointerException
        at IndexHDFS.indexData(IndexHDFS.java:92)
        at IndexHDFS.main(IndexHDFS.java:72)

The problem here is that the destination folder hdfs://hostname:9000/data-Index doesn't exist
and the reason is that when I used copyfromlocal, I embedded data-index/ into multiple folders
(my script did that though!). So the correct form is hdfs://hostname:9000/path/to/data-Index

Now it is fine.Thanks to Sandeep for motivating me for focusing on the address and port.
Regards,
Mahmood




   

  
Mime
View raw message